RESUMEN
OBJECTIVE: This study was designed to establish the reliability of neurologic examination, neuron-specific enolase (NSE), and median nerve somatosensory-evoked potentials (SEPs) to predict poor outcome in patients treated with mild hypothermia after cardiopulmonary resuscitation (CPR). METHODS: This multicenter prospective cohort study included adult comatose patients admitted to the intensive care unit (ICU) after CPR and treated with hypothermia (32-34°C). False-positive rates (FPRs 1 - specificity) with their 95% confidence intervals (CIs) were calculated for pupillary light responses, corneal reflexes, and motor scores 72 hours after CPR; NSE levels at admission, 12 hours after reaching target temperature, and 36 hours and 48 hours after collapse; and SEPs during hypothermia and after rewarming. The primary outcome was poor outcome, defined as death, vegetative state, or severe disability (Glasgow Outcome Scale 1-3) after 6 months. RESULTS: Of 391 patients included, 53% had a poor outcome. Absent pupillary light responses (FPR 1; 95% CI, 0-7) or absent corneal reflexes (FPR 4; 95% CI, 1-13) 72 hours after CPR, and absent SEPs during hypothermia (FPR 3; 95% CI, 1-7) and after rewarming (FPR 0; 95% CI, 0-18) were reliable predictors. Motor scores 72 hours after CPR (FPR 10; 95% CI, 6-16) and NSE levels were not. INTERPRETATION: In patients with persisting coma after CPR and therapeutic hypothermia, use of motor score or NSE, as recommended in current guidelines, could possibly lead to inappropriate withdrawal of treatment. Poor outcomes can reliably be predicted by testing brainstem reflexes 72 hours after CPR and performing SEP.